Alternative management strategies for rheumatoid arthritis (RA) designed to replace the traditional therapeutic pyramid strategy have recently been proposed and are commonly used, even though evidence for their effectiveness is lacking. The most critical clinical question in RA, therefore, is whether these newer strategies are superior to the old in improving long-term functional and quality-of-life outcomes. These important but very difficult questions may be addressed uniquely using ARAMIS data sets which contain multidisciplinary data including disability, pain and patient global assessment outcomes available in some data sets extending over 15 years. This project will address the questions: 1) Is early and consistent Disease-Modifying Anti-Rheumatic Drug (DMARD) use more effective than Non-steroidal Anti-Inflammatory Drug (NSAID) based therapy and, if so, by how much? 2) Are outcomes in RA improving over time? 3) Are clinical courses and response to therapy similar in various ethnic groups, and in Dr4 positive homozygotes, heterozygotes, and Dr4 negative patients? Pilot studies have suggested positive answers to the first two questions. For this project we will improve the quality and quantity of data relevant to these questions, examine and utilize a new national inception cohort patient sample, and employ five complementary analytic approaches: 1) correlation between medication use and outcomes, including disability, pain, global assessment, and x-ray progression, 2) comparisons of temporal trends in medication use and patient outcomes, 3) comparison among matched patients receiving different treatment strategies, 4) simulation of a randomized controlled trial using new econometric instrumental variables, and 5) decision-analytic modelling of alternative strategies. By these means, we hope to guide physicians toward the most effective treatment strategies and to improve patient outcomes in RA.